What does this line of code do when initializing a hash table in python?

Question

I'm looking at the solution for how to create a hash table in python and came across this function:

def __init__(self, size):
    self.size = size
    self.table = [[] for _ in range(self.size)]

The third line of code creating the multi-dimensional array is confusing me. Mainly this part:

for _ in range(self.size)

What exactly is this doing? And why is it needed instead of just doing something like this:

[[] self.size]

Which would create a 2-dimensional array with the size of self.size correct? Any sort of visual aids would really help too.

`[[] self.size]` creates an array whose first element is an empty array, and second element is a number. — Barmar, Mar 27 '17 at 21:49
for the underscore part, see : http://stackoverflow.com/questions/1739514/underscore-as-variable-name-in-python — WNG, Mar 27 '17 at 21:50
`[[] self.size]` is a syntax error. `[[], self.size]` would be a list containing an empty list and whatever `self.size` is. — TigerhawkT3, Mar 27 '17 at 21:53
Okay I was mistaken thinking that `[[] self.size]` was declaring an array of `self.size` size with each index containing an empty array. Thanks for clearing that up. — chadb768, Mar 27 '17 at 21:55
@chadb768 there is no variable declaration in Python. And that isn't an array, that is a list. In fact, it is really a `SyntaxError`, so it's not even that... — juanpa.arrivillaga, Mar 27 '17 at 22:04

Richard · Answer 1 · 2017-03-27T23:05:36.687

1

The line

self.table = [[] for _ in range(self.size)]

creates an array of buckets for holding the contents of the hash table.

The variable name _ is used to indicate that the variable doesn't matter and its contents are, essentially, thrown away. This is useful for unused variables with a short scope, like this one.

You suggest initializing things like this:

self.table = [[]]*self.size

But this is a bad idea because you actually get self.size copies of the same list! That is:

a=[[]]*4
>>> [[], [], [], []]
a[0].append(3)
>>> [[3], [3], [3], [3]]

edited Mar 27 '17 at 23:05

answered Mar 27 '17 at 21:53

Richard

44,865
24
144
216

Of the same *list*. – juanpa.arrivillaga Mar 27 '17 at 22:04
1

@juanpa.arrivillaga: Quite right. I've fixed the mistake. – Richard Mar 27 '17 at 23:05
Yeah, sorry for coming off as pedantic, but it's a pet peeve of mine. Python lists are different than arrays, especially arrays in languages like C++ and Java. They are heterogeneous, resizable array-lists with constant-time append/extend operations. – juanpa.arrivillaga Mar 27 '17 at 23:11
@juanpa.arrivillaga: It was fuzzy wording on my part, and I appreciate the reminder that these are inherently different data structures. – Richard Mar 28 '17 at 00:12

score 1 · Accepted Answer · edited May 23 '17 at 12:25

1

>>> good_table = [[] for _ in range(5)]
>>> good_table
[[], [], [], [], []]
>>> good_table[0].append(3)
>>> good_table
[[3], [], [], [], []]
>>> suggested_table = [[] 5]
  File "<input>", line 1
    suggested_table = [[] 5]
                          ^
SyntaxError: invalid syntax
>>> suggested_table = [[] * 5]
>>> suggested_table
[[]]
>>> [] * 5
[]
>>> bad_table = [[]] * 5
>>> bad_table
[[], [], [], [], []]
>>> bad_table[0].append(3)
>>> bad_table
[[3], [3], [3], [3], [3]]

For the last part, see List of lists changes reflected across sublists unexpectedly.

edited May 23 '17 at 12:25

Community

1
1

answered Mar 27 '17 at 21:56

Alex Hall

31,431
4
39
71

I would love to know why people are downvoting this. What information do I need to add? What hasn't already been thoroughly covered elsewhere? – Alex Hall Mar 27 '17 at 23:55

What does this line of code do when initializing a hash table in python?

2 Answers2